git-send-email: threaded view issue with Thunderbird

Hi,

Sometimes, actually most of the time, git-send-email _seems_ to not work properly with Thunderbird: for example if I send a patchset made up by 4 single patchs, then Thunderbird show me this patchset 'broken' that is [PATCH 2/4] is not shown as a reply to [PATCH 1/4].

For example I can see:

[PATCH 0/4]
       [PATCH 1/4]
[PATCH 2/4]
[PATCH 3/4]
       [PATCH 4/4]

instead of

[PATCH 0/4]
       [PATCH 1/4]
          [PATCH 2/4]
             [PATCH 3/4]
                [PATCH 4/4]

The strange thing is that all email headers are correct.

I recently notice that someone has sent a patchset using git-send-email and I can see it 'broken' as well. The subject of the first email of the set is

[PATCH 0/7] C cleanup series, mostly nasty shadowing

Does anybody have an idea on what's going wrong ?

thanks,

Francis

Le jeu 24 août 2006 10:12, moreau francis a écrit :

here the problem is that I manually sent the 0/7 as the initial mail, and passed its msg-id into --in-reply-to switch of git-send-email. but I did not specified the brackets, and those were not added, meaning that the reference was bad formatted.

I don't know where your other problems come from. What I know for sure is that I will write my own git-send-email version, with less perl, and less broken features, because the two first try I had with git-send-email were quite inconclusive, doing all *but* what I wanted them to do.

if it's not too crude, I will consider submit it to the list as a possible replacement/competitor for git-send-email.

bonjour,

Pierre Habouzit wrote:

well the problem occur with your [PATCH 2/7] and [PATCH 3/7] too. I don't think you manually setup the headers for these patches, do you ?

that would be great since I don't speak perl.

Francis

Pierre Habouzit <madcoder@debian.org> writes:

Please not a revolution but an evolution if possible...

moreau francis <francis_moreau2000@yahoo.fr> writes:

Then this is a wrong list perhaps ;-).

I see that 7+1 series as "broken"; #1 is not a reply to #0. All others are reply to the previous one which is fine:

    Subject: [PATCH 0/7] C cleanup series, mostly nasty shadowing
    Message-ID: <200608231238.10963.madcoder@debian.org>

    Subject: [PATCH 1/7] avoid to use error that shadows the funct
    Message-ID: <11563295562072-git-send-email-madcoder@debian.org>
    In-Reply-To: 200608231238.10963.madcoder@debian.org

    Subject: [PATCH 2/7] git_dir holds pointers to local strings,
    Message-ID: <11563295562422-git-send-email-madcoder@debian.org>
    In-Reply-To: <11563295562072-git-send-email-madcoder@debian.org>

    Subject: [PATCH 3/7] missing 'static' keywords
    Message-ID: <1156329556788-git-send-email-madcoder@debian.org>
    In-Reply-To: <11563295562422-git-send-email-madcoder@debian.org>

    Subject: [PATCH 4/7] remove ugly shadowing of loop indexes in s
    Message-ID: <11563295573215-git-send-email-madcoder@debian.org>
    In-Reply-To: <1156329556788-git-send-email-madcoder@debian.org>

    Subject: [PATCH 5/7] missing #define DEBUG 0 that made the prep
    Message-ID: <11563295573035-git-send-email-madcoder@debian.org>
    In-Reply-To: <11563295573215-git-send-email-madcoder@debian.org>

    Subject: [PATCH 6/7] use name[len] in switch directly, 
    Message-ID: <1156329557424-git-send-email-madcoder@debian.org>
    In-Reply-To: <11563295573035-git-send-email-madcoder@debian.org>

    Subject: [PATCH 7/7] n is in fact unused, and is later shadowed.
    Message-ID: <1156329557613-git-send-email-madcoder@debian.org>
    In-Reply-To: <1156329557424-git-send-email-madcoder@debian.org>

Because #1 has a bogus In-Reply-To field, #0 and #1 would be shown at the same level, but #2-7 would be staggered below #1.

Junio C Hamano wrote:

yeah, that's why I wrote "git-send-email _seems_ to not work properly with Thunderbird". I actually think it's a Thunderbird issue, I'll contact them hoping they won't tell me it's a git issue ;)

thanks

Francis